JP4265374B2 - Image reading apparatus and image processing program - Google Patents

Description

The present invention relates to an image reading apparatus and an image processing program for reading an original shared by a plurality of sensors. In particular, the present invention relates to an image processing technique for improving image unevenness of a read image caused by a difference in characteristics of a plurality of sensors.
The present invention also relates to a data analysis apparatus employed in the image processing technique and a data analysis program thereof.

2. Description of the Related Art Conventionally, an image reading apparatus that scans a document using a plurality of sensors is known (for example, Patent Document 1). This type of image reading apparatus can share an original reading area using a plurality of sensors. As a result, it is possible to greatly shorten the document reading time.
JP 2000-22895 (Claim 1 etc.)

  By the way, some characteristic difference may appear in the plurality of sensors described above. Therefore, conventionally, at the factory shipment stage, etc., adjustments have been made to measure the characteristic differences of a plurality of sensors in a transparent state and remove the measured characteristic differences.

  However, in a situation where the original is actually read, a difference in the original is taken into consideration, and thus a characteristic difference different from the above-described through state appears. As a result, there is a problem in the read image that image unevenness occurs in each read area shared by each sensor, and becomes noticeable as, for example, streak noise or strip noise.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to improve image unevenness of a read image that occurs during document reading.
It is another object of the present invention to provide a data analysis technique suitable for improving such image unevenness.

An image reading apparatus according to a first aspect of the invention scans a document using the plurality of sensors after sharing a reading area of the document in each of a plurality of sensors arranged in parallel in the sub-scanning direction of the document, An image reading apparatus for generating a read image of the original from line data output from each of the plurality of sensors when the original is scanned, wherein a specific area provided on the original is defined by the plurality of sensors. A characteristic difference for each of the plurality of sensors is generated using a histogram statistical unit that obtains a histogram for each of the plurality of sensors using line data obtained by scanning each of the sensors, and the histogram for each of the sensors. A characteristic calculator for obtaining a correction process for correcting the characteristics of the plurality of sensors so as to prevent A gradation conversion unit that reduces image unevenness appearing in the read image by performing the correction process on the line data for each sensor, and the characteristic calculation unit divides the frequency of the histogram into a predetermined ratio. A division analysis unit that obtains frequency division points; and a process determination unit that obtains the correction processing using the correspondence relationship of the frequency division points obtained for each histogram as a conversion rule. The characteristic calculation unit includes: According to at least one of the type and brightness, at least one of “the number of frequency division points” and “the value of the predetermined ratio” is changed.

According to a second aspect, in the first aspect, the characteristic calculation unit determines that the frequency division is performed when the original is a negative (an original in which the brightness of an image is reversed) or when the original is evaluated to be dark. The number of points is increased .
The conditions “negative” and “dark original” are common in that the image unevenness of the read image is conspicuous.

According to a third invention, in the first or second invention, the characteristic calculation unit stores the past correction process, and performs a weighted average of the current correction process and the past correction process to obtain line data of the read image. It is characterized in that a correction process to be performed is obtained.

In a fourth aspect based on any one of the first to third aspects, the histogram statistical unit compares the line data for each sensor in a pre-scan prior to the main scan and determines the necessity of the correction processing. It is characterized by that.

In a fifth aspect based on any one of the first to fourth aspects, the histogram statistics unit and the calculation unit obtain the correction processing by a pre-scan having a lower resolution than the main scan, and the gradation conversion unit The correction process is performed on the line data of the main scan .

In a sixth aspect based on the fifth aspect, the plurality of sensors photoelectrically convert a common portion of the document in the pre-scan, and the histogram statistical unit is configured to detect each sensor from line data of the common portion. A histogram is obtained, and the characteristic calculation unit obtains the correction processing based on a difference in the histogram at the common location .

A seventh invention is the difference between the first and sixth inventions, wherein the pre-scan is performed prior to the main scan, and the evaluation unit that evaluates the difference in line data between the sensors, and the difference in the line data And a main scanning unit that performs the main scanning of the original document by reducing the number of sensors used when the evaluation is determined to be outside the predetermined allowable range .

An image processing program according to an eighth invention causes a computer to function as the histogram statistical unit, the characteristic calculation unit, and the gradation conversion unit according to any one of claims 1 to 9. .

  An image reading apparatus according to a ninth aspect of the invention scans the target using the plurality of sensors after sharing the target reading area in each of the plurality of sensors arranged in parallel in the target sub-scanning direction, An image reading apparatus that generates a read image of the target from line data output from each of the plurality of sensors when the target is scanned, wherein a specific region provided on the target is defined by the plurality of sensors. A characteristic difference for each of the plurality of sensors is generated by using a histogram statistical unit that obtains a histogram for each of the plurality of sensors using line data obtained by scanning each, and the histogram for each of the sensors. A characteristic calculator for obtaining a correction process for correcting the characteristics of the plurality of sensors so as to prevent A gradation conversion unit that reduces image unevenness appearing in the read image by performing the correction process on the line data for each sensor, and the characteristic calculation unit divides the frequency of the histogram into a predetermined ratio. A division analysis unit that obtains a frequency division point; and a process determination unit that obtains the correction process using a correspondence relationship between the frequency division points obtained for each histogram as a conversion rule. According to the evaluation determination result of whether or not the image is dark, at least one of “the number of frequency division points” and “the value of the predetermined ratio” is changed.

As described above, in the present invention, a correction process for correcting the characteristics of a plurality of sensors is obtained from a histogram for each sensor so as not to cause a difference in sensor characteristics . According to this correction processing creation rule, the influence of noise is small because it is handled collectively as a frequency, and it is possible to obtain a high-quality correction processing in order to equalize the characteristic differences.
In particular, in a multi-sensor type image reading apparatus, by performing this correction process, image unevenness due to a characteristic difference between sensors can be made inconspicuous.

Hereinafter, embodiments of an image reading apparatus (including a data analysis apparatus) according to the present invention will be described with reference to the drawings.
FIG. 1 is a diagram illustrating the configuration of the image reading apparatus 11.
In FIG. 1, a film 12 is inserted into the image reading device 11 as a document from the outside. The film 12 is irradiated with illumination light (R light, G light, B light, IR light, etc.) sequentially from the illumination unit 13. At this time, the light transmitted through the film 12 forms a light image by an imaging lens (not shown). A sensor block 15 is disposed on the image plane of the optical image. In the sensor block 15, a plurality of sensors 15 </ b> A to 15 </ b> C (here, three as an example) are arranged at intervals in the scanning (sub-scanning) direction of the film 12.

Further, the image reading apparatus 11 is provided with a scanning unit 14. The scanning unit 14 reads line data from the sensors 15A to 15C while relatively moving the film 12 and the sensor block 15 in the scanning direction.
The line data output from the sensors 15 </ b> A to 15 </ b> C passes through the analog signal processing circuit 16 and then is input to the A / D conversion unit 17. The A / D conversion unit 17 performs digital conversion while switching these line data in a time-sharing manner using an analog switch or the like.

The line data digitized by the A / D conversion unit 17 is subjected to dark voltage correction and shading correction via the correction unit 17a. The line data output from the correction unit 17a is output to the memory 18 and the gradation conversion unit 19. The gradation conversion unit 19 includes gradation conversion tables (LUT1 to 3) for each of the sensors 15A to 15C and for each color. The line data subjected to gradation correction by the gradation conversion unit 19 is output to an external computer via the interface unit 20.
In addition, the image reading apparatus 11 is provided with an MPU (Micro Processor Unit) 21 for system control.

[Correspondence with Invention]
The correspondence relationship between the invention and this embodiment will be described below. Note that the correspondence relationship here illustrates one interpretation for reference, and does not limit the present invention.
The histogram statistic unit described in the claims corresponds to the “function of performing a pre-scan and obtaining a histogram of line data for each of the sensors 15A to 15C” of the scan unit 14 and the MPU 21.
The characteristic calculation unit described in the claims corresponds to “a function for obtaining gradation correction for aligning corresponding histograms and setting it in the gradation conversion table of the gradation conversion unit 19” of the MPU 21.
The gradation converter described in the claims corresponds to the gradation converter 19.
The division analysis unit described in the claims corresponds to the “function for obtaining the frequency division point from the histogram” of the MPU 21.
The processing determination unit described in the claims corresponds to the “function for obtaining gradation correction according to the conversion rule for associating the frequency division points between the sensors 15 </ b> A to 15 </ b> C” of the MPU 21.
The evaluation unit described in the claims corresponds to the “function for evaluating the output difference of the sensors 15A to 15C in the pre-scan” of the MPU 21.
The main scanning unit described in the claims corresponds to the “function for setting and changing the number of sensors 15A to 15C used for the main scanning based on the evaluation result” of the scanning unit 14 and the MPU 21.

[Description of operation of this embodiment]
2 and 3 are flowcharts for explaining the operation of the image reading apparatus 11.
Hereinafter, the characteristic operation of this embodiment will be described with reference to this flowchart.

Step S <b> 1: First, the user inserts the film 12 into the image reading device 11. At this time, the sensor block 15 is located at a through position of the film 12. The MPU 21 reads the dark voltage from the dark sensors 15 </ b> A to 15 </ b> C while maintaining the illumination unit 13 in the off state. These dark voltages are respectively held in the correction unit 17a and used for dark voltage correction described later.
Subsequently, the MPU 21 sequentially turns on the illumination unit 13 in each color. At this time, the sensors 15A to 15C output the shading change of each color. The correction unit 17a stores these shading changes. The stored shading change is used for shading correction described later.
Further, the MPU 21 gives a linear characteristic as a default setting to all the gradation conversion tables of the gradation converting unit 19.

Step S2: In this state, the MPU 21 enters a preview instruction standby state.
When the user gives a preview instruction from the external computer, the MPU 21 shifts the operation to step S3.

Step S3: The scanning unit 14 scans the film 12 at a high speed, and determines an appropriate exposure amount (such as illumination time) for each color from the output signal level of the high speed scan.

Step S4: Using the exposure amount of each color determined in step S3 (the same exposure amount as the main scan described later), the scanning unit 14 performs a low-resolution pre-scan.
At this time, the scanning unit 14 controls the sensor block 15 so that each of the sensors 15A to 15C reads a common portion of the film 12.
For example, such pre-scanning can be realized by alternately executing scanning of the sensor block 15 and line reading in a moving cycle in which the intervals of the sensors 15A to 15C are equally divided.

Further, for example, the common locations roughly set on the film 12 may be read by the sensors 15A to 15C, and the other locations may be read by the sensors 15A to 15C. Such a control sequence can speed up the pre-scan.
The line data for each sensor and each color generated by the pre-scan is subjected to dark current correction and shading correction by the correction unit 17a, and then temporarily stored in the memory 18.
These line data are output to an external computer via the interface unit 20 after being subjected to gradation conversion for preview in the gradation conversion unit 19.

Step S5: The external computer generates a preview image from the pre-scan line data and displays it on the monitor screen. The user can change the reading range during the main scan by performing a GUI operation while viewing the preview image.
This reading range change instruction is transmitted to the MPU 21 via the interface unit 20. The MPU 21 matches the read range and limits the use range of the line data held in the memory 18.

Step S6: In this state, the MPU 21 enters a standby state for the main scan instruction.
When the external computer gives a main scan instruction via the interface unit 20, the MPU 21 shifts the operation to step S7.

Step S7: The MPU 21 reads the line data in the memory 18, and obtains the difference of the line data for each sensor and for each color. Based on the difference of the line data, the characteristic difference of the sensors 15A to 15C that are surfaced during the main scan of the film 12 is predicted.

Step S8: The MPU 21 determines whether or not the characteristic difference between the sensors 15A to 15C is within a predetermined allowable range.
Here, when it falls within the allowable range, the MPU 21 shifts the operation to step S17 (main scan).
On the other hand, when it is outside the allowable range, the MPU 21 shifts the operation to step S9.

Step S9: The MPU 21 selects a reference sensor (hereinafter referred to as “reference sensor”) from the line data in the memory 18.
For example, this selection is realized by obtaining a representative value such as an intermediate value or an average value for line data at a common location, and selecting line data closest to the representative value.
Next, the MPU 21 determines a threshold value of the difference in line data between the reference sensor and other sensors, and determines which color output of which sensor needs gradation correction.

Step S10: Here, the MPU 21 reads the user setting, and determines the low-speed scan permission setting (setting whether or not the number of sensors 15A to 15C can be used).
If the low speed scan is permitted, the MPU 21 shifts the operation to step S11.
On the other hand, when the low-speed scan is not permitted, the MPU 21 shifts the operation to step S12.

Step S11: The MPU 21 disables the sensor that requires gradation correction, and shifts the operation to Step S17 (main scan).
In the main scan in this case, the film 12 is read using only a sensor whose characteristic difference falls within a threshold value.

Step S12: Thereafter, correction processing for correcting the characteristic difference between the sensors 15A to 15C is determined.
First, the MPU 21 reads the pre-scan line data from the memory 18 and obtains a histogram of the line data for each sensor and for each color. These histograms are held in the memory 18 as statistical data.

Step S13: Next, the MPU 21 determines whether or not the film 12 is a negative film. Subsequently, the MPU 21 evaluates whether or not the document is a dark document by determining the threshold value of the brightness level of the line data.
Here, if the reading range of the film 12 is a negative film or a dark original, the MPU 21 shifts the operation to step S14.
On the other hand, in other cases, the MPU 21 shifts the operation to step S15.

Step S14: When the film 12 is a negative film or a dark original, image unevenness due to a characteristic difference between the sensors 15A to 15C is easily noticeable. Therefore, in order to obtain a more precise correction process, the MPU 21 increases the number of calculated frequency division points. At this time, it is preferable to concentrate a predetermined ratio described later on the dark side so that the frequency division points are dense on the dark side of the histogram.
After such a setting change, the MPU 21 shifts the operation to step S15.

Step S15: The MPU 21 divides the histogram for each sensor and each color into a plurality of predetermined ratios as shown in FIG.
The frequency division points shown in FIG. 4 are points at which the cumulative frequency from the dark side reaches a predetermined ratio (for example, 0.42%, 10%, 30%, 50%, 70%, 90%, 97%).
On the other hand, in the case where the setting is changed in step S14, 20%, 40%, 60%, and 80% are added to the predetermined ratio.
The darkest power division point is preferably set to about 0.42% from the dark side in order to reduce the influence of power concentration and dark noise due to black crushing.
The brightest frequency division point is preferably set to about 97% from the dark side in order to reduce the influence of frequency concentration due to saturation (out-of-white).

Step S16: As shown in FIG. 5, the MPU 21 obtains a polygonal line correction process for converting the frequency division point to be corrected into the frequency division point of the reference sensor. In addition, about the both ends of this broken line, the broken line of the darkest and brightest frequency division point is extended.
The MPU 21 obtains such correction processing for each color and sets it in the LUT of the gradation conversion unit 19.

Step S17: The MPU 21 instructs the scanning unit 14 to perform a main scan of the film 12. In this main scan, the reading operation of the film 12 is shared by the sensors 15A to 15C, whereby the reading operation is completed with high definition and at high speed.

Step S18: The line data of the main scan is subjected to correction processing in the gradation conversion unit 19 according to the LUT for each sensor and each color.
The main scan line data processed in this way is output to an external computer via the interface unit 20.
The external computer combines the line data to generate a read image of the film 12.

[Effects of this embodiment, etc.]
As described above, in the present embodiment, correction processing is obtained according to the film 12, and image unevenness due to the characteristic difference between the sensors 15A to 15C is reduced. Therefore, in consideration of the difference of the film 12, a characteristic difference different from the above-described through state can be detected. Therefore, it is possible to reliably reduce the image unevenness that appears intricately related to the solid difference of the film 12 and the difference between the sensors 15A to 15C.

  By the way, this type of correction processing can also be obtained from a conversion rule for associating sensor-specific pixel values. However, since the pixel value here includes noise, it is difficult to obtain matching consistency, and it is difficult to obtain a high-quality correction process. In particular, a dark manuscript is greatly affected by noise, and a practical correction process cannot be obtained.

  On the other hand, in the present embodiment, correction processing is obtained in a direction in which the histograms for each sensor are substantially matched. In the process of creating such correction processing, the pixel values are collectively handled in units of frequency, so that the influence of noise included in each pixel value is greatly reduced. As a result, an extremely high quality correction process can be obtained.

  In particular, in a dark original, the noise tendency for each sensor also appears in the histogram. In this case, in the creation of the correction process based on the above-described histogram, it is possible to obtain a correction process that substantially matches the noise tendency for each sensor without being confused by the instantaneous noise value. As a result, it is possible to remove even the image unevenness of the noise tendency generated in the read image. This is an effect that cannot be realized by association of pixel values.

  Furthermore, in the present embodiment, a correction process for making the histograms substantially coincide with each other is strictly obtained by a data analysis technique focusing on the frequency division points. That is, in the present embodiment, first, the frequency of the histogram (X shown in FIG. 5) is divided into a predetermined ratio to obtain the frequency division point. This frequency division point corresponds to a numerical value on the variable axis of the histogram. Next, as shown in FIG. 5, a correction process according to a conversion rule for converting a frequency division point to be corrected into a reference frequency division point is obtained.

  By applying the correction process obtained in this way to the correction target data, the histogram of the corrected data can be substantially matched with the reference histogram (Y shown in FIG. 5). By such a data analysis technique, it has become possible to surely and carefully correct a very fine image unevenness (characteristic difference) to the extent that the surface is formed by the difference in the film 12.

  Furthermore, in this embodiment, the number of frequency division points and the value of the predetermined ratio are changed according to the type and brightness of the film 12. Therefore, an appropriate correction process can be obtained according to the situation of the film 12.

In particular, in this embodiment, in the case of a negative or dark original, the frequency division points are arranged at a close interval. As a result, more precise correction processing can be applied to image unevenness that is particularly noticeable in negatives and dark originals.
In the present embodiment, a pre-scan is performed prior to the main scan. In this preliminary scan, the necessity for correction processing is determined by comparing line data. According to this necessity determination, correction processing can be efficiently omitted for the color signal of the sensor that does not require correction processing.

  Furthermore, in this embodiment, this pre-scan is performed at a lower resolution than the main scan, and correction processing is obtained from the result of the pre-scan. In this case, it is possible to significantly reduce the processing time and memory capacity required for calculating the correction process, rather than obtaining the correction process from the result of the high-resolution main scan. This is particularly advantageous when the correction processing characteristic calculation is implemented in hardware.

  Moreover, in this embodiment, in this prior scan, the common location of the film 12 is read using sensors 15A-C, respectively. Since the correction process is obtained from the line data of the common portion, it is possible to obtain a more accurate correction process without being affected by the difference in the design.

  Furthermore, in the present embodiment, when the limitation on the number of sensors 15A to 15C used is permitted, the sensor that needs correction processing is disabled and the main scan is performed. Such an operation can also improve image unevenness due to a characteristic difference between sensors.

[Supplementary items of this embodiment]
In the above-described embodiment, the color image is read by switching the color of the illumination light. However, the present invention is not limited to this. For example, a color image reading may be realized by arranging a color filter array in the sensor. Also in this case, the correction process can be obtained based on the histogram as in the above-described embodiment. In the correction process in this case, the characteristic difference of the color filter array can also be suppressed.

  Further, in the above-described embodiment, as shown in FIG. 5, a correction process using a linear characteristic as a standard is obtained. However, the present invention is not limited to this. For example, gradation conversion such as gamma conversion and bit number conversion can be performed in combination with the correction processing of this embodiment.

  In the embodiment described above, a histogram having (pixel value, frequency) as an axis is obtained as shown in FIG. However, the present invention is not limited to this. For example, a histogram with (pixel value, cumulative frequency from the dark side) as an axis may be obtained. In this case, it is possible to immediately obtain the pixel value corresponding to the frequency division point by referring to this histogram for the dark side cumulative frequency indicated by the predetermined ratio.

  In the above-described embodiment, the correction process obtained in step S16 is set in the LUT of the gradation conversion unit 19 as it is. However, the present invention is not limited to this. For example, a past (for example, previous time) correction process may be weighted and averaged in the latest correction process, and the correction process after the weighted average may be set in the LUT of the gradation conversion unit 19. In this case, there is no fear that the correction process fluctuates greatly every time scanning is performed, and the continuity and stability of the correction process are guaranteed.

  Further, in the above-described embodiment, the correction process is obtained for each frame of the film 12. However, the present invention is not limited to this. For example, the correction process may be obtained for each sleeve or each film. Further, for example, when the “image difference” and the “characteristic difference between the sensors 15 </ b> A to 15 </ b> C” affect the image unevenly, the image region is appropriately divided and a correction process is obtained for each divided region. May be.

  In the above-described embodiment, as shown in FIG. 5, a polygonal line correction process using linear interpolation is obtained. However, the present invention is not limited to this. For example, you may obtain | require the correction process which carried out the curve interpolation of the broken line shown in FIG.

Further, in the above-described embodiment, the frequency division points of the histogram are respectively obtained after the predetermined ratio is first determined. However, the present invention is not limited to this. For example, the correction processing may be obtained in the following order.
(1) First, a desired frequency division point is determined for one histogram.
(2) The frequency of one histogram is divided at these frequency division points to obtain a predetermined ratio value.
(3) The frequency division point is obtained by applying the obtained predetermined ratio to the other histogram.
(4) A correction process is obtained according to a conversion rule that associates the frequency division points of both histograms.

  Further, in the above-described embodiment, the case where the present invention is realized by the image reading device 11 (firmware of the MPU 21) has been described. However, the present invention is not limited to this. For example, all or a part may be implemented by hardware.

  For example, the present invention may be implemented in a driver program or an image processing program on the external computer side.

  Further, for example, it is of course possible to provide the correction processing of the present invention through a communication line such as the Internet.

  Further, for example, the correction processing of the present invention may be realized by a flood bed scanner, a handy scanner, or a DPE device with a film scanner.

  In particular, the correction processing technique for aligning the histogram with the reference histogram can be applied to various purposes as a general-purpose technique for adjusting input data to a desired histogram.

  As described above, the present invention is a technique that can be used for an image reading apparatus, an image processing program, a data analysis apparatus, a data analysis program, and the like.

2 is a diagram illustrating a configuration of an image reading device 11. FIG. 3 is a flowchart (first half) for explaining the operation of the image reading apparatus 11; 4 is a flowchart (second half) for explaining the operation of the image reading apparatus 11; It is a figure explaining a histogram and a frequency division point. It is a figure which shows an example of a correction process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Image reader 12 Film 13 Illumination part 14 Scan part 15 Sensor block 15A Linear sensor 15A-C Sensor 16 Signal processing circuit 17 A / D conversion part 17a Correction part 18 Memory 19 Tone conversion part 20 Interface part 21 MPU

Claims (9)

The document is scanned using the plurality of sensors after sharing the reading area of the document in each of a plurality of sensors arranged in parallel in the sub-scanning direction of the document, and the plurality of sensors are scanned when the document is scanned. An image reading device that generates a read image of the original from line data output from each of the sensors ,
A histogram statistical unit that obtains a histogram for each of the plurality of sensors using line data obtained when each of the plurality of sensors scans a specific area provided on the document;
By using the histogram for each of the sensors, the characteristic calculating section for obtaining a correction process for correcting the characteristics of the plurality of sensors so that the characteristic difference of each of the plurality of sensors does not occur,
A gradation converter that reduces image unevenness appearing in the read image by performing the correction process on the line data for each sensor constituting the read image ; and
The characteristic calculation unit obtains the correction processing using a division analysis unit that obtains a frequency division point by dividing the frequency of the histogram into a predetermined ratio, and a correspondence between the frequency division points obtained for each histogram as a conversion rule. A processing determination unit,
The characteristic calculation unit changes at least one of the “number of frequency division points” and the “predetermined ratio value” according to at least one of the type and brightness of the document. . The image reading apparatus according to claim 1,
The characteristic calculation unit increases the number of frequency division points when the original is a negative (original with reversed image brightness) or when the original is evaluated to be dark. apparatus. The image reading apparatus according to claim 1 or 2,
The image reading apparatus is characterized in that the characteristic calculation unit stores the past correction processing and obtains correction processing to be applied to the line data of the read image by performing weighted averaging of the current and past correction processing . The image reading apparatus according to any one of claims 1 to 3,
The image processing apparatus according to claim 1, wherein the histogram statistics unit compares the line data for each sensor in a pre-scan prior to the main scan and determines the necessity of the correction process . The image reading apparatus according to claim 1, wherein:
The histogram statistics unit and the calculation unit obtain the correction process by a pre-scan with a lower resolution than the main scan,
The image reading apparatus , wherein the gradation converting unit performs the correction process on the line data of the main scan . The image reading apparatus according to claim 5 .
The plurality of sensors photoelectrically convert a common portion of the document in the preliminary scan,
The histogram statistics unit obtains a histogram for each sensor from the line data of the common part,
The image reading apparatus , wherein the characteristic calculation unit obtains the correction processing based on a difference in the histogram at the common portion . The image reading apparatus according to any one of claims 1 to 6,
Prior to the main scan, a pre-scan is performed, and an evaluation unit that evaluates and determines a difference in line data between the sensors;
When the difference in the line data is evaluated and determined to be outside a predetermined allowable range, a main scanning unit that performs the main scanning of the document by reducing the number of sensors used;
Image reading apparatus characterized by comprising a. An image processing program for causing a computer to function as the histogram statistic unit, the characteristic calculation unit, and the gradation conversion unit according to any one of claims 1 to 7 . The target is scanned using the plurality of sensors after sharing the target reading area in each of the plurality of sensors arranged in parallel in the target sub-scanning direction, and the plurality of sensors are scanned when the target is scanned. An image reading device that generates a read image of the object from line data output from each of the sensors,
A histogram statistical unit that obtains a histogram for each of the plurality of sensors, using line data obtained when each of the plurality of sensors scans a specific area provided in the target;
Using the histogram for each sensor, a characteristic calculation unit for obtaining a correction process for correcting the characteristics of the plurality of sensors so as not to cause a characteristic difference for each of the plurality of sensors;
A gradation converter that reduces image unevenness appearing in the read image by performing the correction process on the line data for each sensor constituting the read image; and
The characteristic calculation unit obtains the correction process using a division analysis unit that obtains a frequency division point by dividing the frequency of the histogram into a predetermined ratio, and a correspondence between the frequency division points obtained for each histogram as a conversion rule. A processing determination unit,
The characteristic calculation unit changes at least one of “the number of frequency division points” and “the value of the predetermined ratio” according to an evaluation determination result as to whether or not the target is dark.
An image reading apparatus.

Images